Study On The Volatility Of China’s CSI 300 Stock Index Futures Based On MEM Model

Economic globalization is greatly increasing the efficiency of the global financial markets. Financial markets of all countries constitute a mutual financial system. While capital flows across markets, financial risk also moves quickly. So the top issue of developing financial markets in the world is risk prevention, and the core issues of risk prevention are the precise measurement and accurate prediction of risk.This paper focuses on the CSI 300 stock index futures market in China. Based on the Multiplicative Error Model(MEM) we studied the risk, including the volatility measurement, prediction, value at risk(VaR) prediction and the determination of dynamic hedging ratio. The volatility measurement we use is Weighted Realized Range-based volatility(WRRV) which is based on high frequency data. We assume that the residual terms of the MEM model follow three kinds of distribution, namely standard exponential distribution, standard Weibull distribution and standardized generalized Gamma distribution.We studied the out-of-sample forecast by using MEM, and based on the forecast we studied the value at risk(VaR) prediction and dynamic hedging ratio determination of the CSI 300 index futures. In predicting value at risk(VaR), we use the variance-covariance analysis method and peaks over threshold(POT) model respectively to combine with GARCH model, EGARCH model, MEM and then measure VaR. Use the Kupiec LR test to measure model effectiveness. The empirical results showed that the POT model is better than variance-covariance analysis method. In the framework of POT model, comparing the MEM model with GARCH model, we find that MEM-EVT-POT class models are superior to GARCH-EVT-POT class models. In determining the dynamic hedging ratio, we introduce the Copula method to determine the Kendall rank correlation coefficient between stock index futures and spot markets, and respectively use the GARCH model and MEM to predict conditional volatility. The empirical result showed that the hedging ratio based on Copula-MEM is better than that of Copula-GARCH model.